My research project during my PhD is to balance human and user efforts for data visualizations. One of the methods to balance this effort is to create systems that automate the process of visualization design (e.g., visualization recommendation systems). There are different ways to create underlying models for recommending visualizations. For instance, many of the existing visualization recommendation tools are rule-based. However, I am sure rule-based models are not the only way. So, I decided to take a course called “introduction to cognitive science” this semester. I am excited to explore different models that we can build in cognitive science and can be useful for my future visualization recommendation system.
At the high level, in this class we want to create an agent that models how human’s brain process for a specific task. For example, suppose we want to create an agent for finding a friend. The task is finding a friend. We first need to observe and extract factors that a human normally does to find a friend. After extracting all the factors, we need to model this. While I am still not sure how we are going extract these factors and model them, based on the class discussion it seems that we need to do some sort of surveys or interviews to collect data. We then need to extract those factors from the collected data.